Navigate the websites work runs on
Open pages, follow links, scroll, and work through multi-step websites. Appaca can use the screen when the task needs more than reading a page.
Appaca can browse websites, fill in forms, and work with files on its own computer. Every conversation gets an isolated desktop. Watch it work, or take control when you need.
I’ll open the supplier portal, fill in your requirements, and save the enquiry as a draft.
Your enquiry for 24 ergonomic office chairs, delivered to Melbourne, is saved as a draft. It’s ready for you to review in the computer.
Some work needs a browser and a few well-placed clicks. Give Appaca the outcome and the details. It can work through the interface, one step at a time.
Open pages, follow links, scroll, and work through multi-step websites. Appaca can use the screen when the task needs more than reading a page.
Give Appaca the details to enter. It can fill in forms and work with web apps through their interface, including sites without an API.
Download files and use the desktop’s file manager, text editor, terminal, and office apps as part of the task.
A supplier enquiry in one conversation. Research in another. Each conversation has its own isolated computer, so the desktop for one stays separate from the next.
Within a conversation, Appaca can pick up from the previous step. The browser and desktop keep their state between computer-use calls.
Open the computer from your conversation to see what’s happening. Take control to sign in, adjust a detail, or handle a step yourself.
The agent waits while you use the computer. Sign in directly on the desktop when a website needs your account.
Leave a note about what you changed, then hand control back. Appaca looks at the screen and continues from there.
More about computer use, browser tasks, and staying in control.
Computer use gives Appaca’s AI agent an isolated virtual Linux desktop with a web browser, terminal, file manager, text editor, and office apps. The agent looks at screenshots and uses actions such as clicking, typing, scrolling, and keyboard shortcuts to carry out your instructions.
Appaca can navigate websites, fill in forms, interact with web apps, and download files. Computer use is useful for multi-step work that needs a screen, including websites without an API. Some sites may require you to sign in or complete a step yourself. Whether a task can be completed depends on the website and the access available.
Yes. Each conversation has its own isolated computer, separate from the computers used in other conversations. The agent starts or resumes that computer when a task needs it, and can continue from its existing state between computer-use steps in the conversation.
Yes. Open Computer in the conversation’s side panel to watch the desktop. Choose Take control to use it yourself; the agent waits while you are in control. When you are ready, add an optional note and choose Hand back so the agent can continue.
The agent asks you to take control and sign in directly on the virtual computer. Passwords, one-time codes, and payment details should not be included in your task instructions. Once you have finished the step, hand control back to Appaca.
The computer keeps its state between computer-use steps and can resume after being paused. You can also pause or reset it from the computer options. Computers may be cleaned up after an extended period of inactivity, and resetting clears the current environment, so do not treat the desktop as permanent storage.
Yes. Computer use draws on your workspace’s AI credits for the agent’s AI usage and the time the computer is running. It pauses after a period of inactivity to save resources. The credits used depend on the task and its runtime.
Let super intelligence handle the browser tasks and everyday admin, so you can focus on what needs you.